Introducing Topicalizability
Let me present the full paradigm relevant for the subsequent discussion. As Beck (1996a) notes, many elements may occur in between two wh phrases or follow them. (The examples presented in this section are variations of constructions first discussed in Grohmann 2000a, 2003a.)
This is not so for all quantificational phrases, though. Considering both ordering options for quantificational elements, we can observe a stark contrast (see (2)): (some) universal quantifiers are able to appear in between the two wh-elements, while negative ones are not. If all wh-elements really must end up in CP at one point (through wh-movement at LF, for example), Beck’s barrier approach would provide one possible account for these facts. Note that nothing rules out the co-occurrence of two wh elements and a negative quantifier: if the apparently intervening quantifier follows all wh-elements, the question is well formed, as the examples in (1) and (2) illustrate.
(1)

(2)

However, not only negative quantifiers are prohibited from intervening, nor is the class of possible interveners restricted to universal quantifiers.
(3)

And again, if the quantifiers follow the wh-items, everything is all right.
(4)

These contrasts seem to allow for the following generalization:
(5)

Now, this generalization seems to fit the characterizations of intervention presented earlier rather well—but does that make it any more plausible? I do not think so. Rather, I believe, this state of affairs begs to be explained by something deeper, preferably by being tied to other properties of the grammar.
Let me introduce and investigate further one such property: topicalizability, the property of being able to serve as a topic. As it turns out, there is evidence for the assumption that (monotone-) decreasing quantifiers are nontopicalizable.
Consider the contrast in (6)–(7), where capitals indicate focus.
(6)


(7)

Next, consider the following questions
(8)

For a felicitous answer to (8a–b) in German, the new information appears in what we may call a default focus position, where the type of focus is information focus (as summarized in E ´. Kiss 1998) and the de fault position is rather low in the structure (see Meinunger 2000 for useful discussion and references).
In replying to (8b), neither construction in (6) or (7) can be used; the only word order that results in a felicitous answer is that in (9).
(9)

However, (6a) and (7a) can be used as felicitous answers to questions like (8c–d), distinguishing them from the context, where the answer contains a contrastive element as part of the new information. (6b) and (7b) are simple topic structures: there is no contrastive stress, and the fronted elements can only pick up a referent already established in the discourse; both are infelicitous.
Let me propose a final argument against any intervention-based account of the ‘‘Beck effect’’ (which we can now understand as the phenomenon of specific word order incompatibilities between nontopicalizable elements and the two wh-phrases in German multiple wh-questions). We can construct paradigms that are identical to the ones discussed above, except that they do not contain any quantificational element. If it can be shown that here, too, we find a topicalizability effect, (any type of) intervention is ruled out for good.